Longitudinal Evaluation of Optic Disc Measurement Variability With Optical Coherence Tomography and Confocal Scanning Laser Ophthalmoscopy

Abstract

To evaluate and compare the longitudinal variability of optical coherence tomography (OCT) and confocal scanning laser ophthalmoscopy (CSLO) optic disc measurements. A total of 25 normal and 50 glaucomatous eyes from 75 subjects were included in the analysis. The optic disc was measured by OCT and CSLO. Three separate measurements collected over an average period of 8.5+/-0.9 months were used to evaluate reproducibility. Univariate and multivariate regression analyses were performed to evaluate the associations between age, refraction, diagnosis (glaucoma or normal), visual field mean deviation, optic disc area, signal strength variance (OCT), optic disc area variance (OCT), image quality SD (CSLO), reference height variance (CSLO), and rim area variability. The intraclass correlation coefficient of optic disc measurements (except for optic disc area) ranged between 0.86 and 0.95 for OCT and between 0.89 and 0.96 for CSLO. The intraclass correlation coefficient for rim area measurement was significantly higher in CSLO (0.95) than that of OCT (0.86, P<0.001). After adjustment for other predictors, optic disc area variance and reference height variance were the most important factors responsible for rim area variability in OCT and CSLO, respectively. Although both OCT and CSLO have relatively low variability for optic disc measurements, CSLO demonstrates higher measurement reproducibility for rim area compared with OCT. Variations of disc area in OCT and reference height in CSLO constituted a significant proportion of the rim area variability during longitudinal assessment.